How AI's Aiding Microsoft Map England's Carbon-Rich Peatland

With scientists warning about climate change posing an ever greater threat with each passing year, enterprises and organisations alike are having to take a multipronged approach to learn how they can apply their abilities to tackle it.

Alongside carbon capture technologies, or optimising flightpaths to reduce energy, one focus is on natural UK’s peatlands.

The peatlands covering 12% of England country, through their role as vital carbon sinks, lock in an estimated 3.2 billion tonnes of carbon.

Yet, their existence is under threat, and so Microsoft, in collaboration with the UK government's Department for Environment Food and Rural Affairs, is working on a solution with AI. 

Mapping peatland damage

Only 22% of English peatlands remain in a near-natural state. Therefore, Microsoft and Defra created Natural England’s innovative AI for Peat programme, which tackles this issue by restoring these degraded ecosystems using advanced AI.

Restoration of peatlands begins with understanding their condition, a process that traditionally relied on slow and laborious manual surveys.

“We can’t start to restore them until we know their current condition and where any damage is located,” says Karen Rogers, Lead Advisor in Natural England’s Cheshire and Lancashire Area Team.

Degraded peatlands, which suffer from issues such as overgrazing, drainage, and pollution, are at risk of releasing large amounts of carbon into the atmosphere. The scale of this threat is considerable; it is estimated that degraded peatlands in England could emit as much as 584 million tonnes of carbon.

Given the pressing nature of this challenge, Natural England has turned to AI in its quest for a more effective and integrated mapping solution.

AI's ability to map peatlands

In 2021, an initiative proposing AI-based mapping secured initial funding through the Civil Service Data Challenge, setting the stage for the AI for Peat project.

Utilising deep learning models on Microsoft Azure and Azure Databricks, this project seeks to detect peatland characteristics, including grips (drainage channels) and gullies, by analysing high-resolution aerial imagery.

“This is life-saving work, so why wouldn’t we support it with leading technology?,” asks Joe Hillier, Head of the Analytics Directorate at Natural England. “We’re at the cusp of AI’s potential and projects like this exemplify that.”

Nonetheless, the obstacles are considerable. Processing aerial imagery at a resolution of 12.5 centimetres per pixel demands significant computational resources. The Data and Analytics Science Hub (DASH), established by Defra in 2022, has provided the necessary cloud infrastructure.

Olivia Newport, Head of DASH, says: “We want to bring the analytics and science community together, supporting them with scalable compute and innovation.”

National mapping and local restoration

The incorporation of AI into peatland mapping has revolutionised the restoration efforts. Rather than spending years on manual mapping, local teams can now leverage AI-generated data to pinpoint areas that need attention. 

“We can load the output mapping layers onto a tablet and follow the features in the field,” Karen explains. “It makes everything so much easier. Instead of mapping features, we’re asking ‘What intervention should I be making here?’”

Although this technology represents a significant advancement, challenges persist. Variations in vegetation and imagery can affect the precision of AI models. “We are currently exploring how to ensure that these models can be effectively applied throughout England,” explains Martha Tabor, Data Scientist at Defra.

Despite these hurdles, the programme has huge potential. “If we can add carbon metrics to the mapping, we’ll get really useful digital datasets,” says Anne Williams, Project Manager of Natural England's AI for Peat.

She envisions the development of a digital twin to model the impacts of restoration interventions, providing critical insights for future projects.

Foundation for future collaboration

With the AI for Peat project involving numerous players, each are bringing skills to help the Peatland restoration efforts. 

“Accepting that no one gets things right first time is really important. Being willing to take risks helps push the boundaries of what’s possible," says Paul Sinclair, Head of Data Exploitation at Defra.

For many involved, the significance of the work goes beyond technical innovation.

Although they are now accelerating, the real takeaway is the adoption of UK government agencies adopting AI across its departments.

AI's work in environmental endeavours is only beginning, yet efforts like this highlight its impact to make serious efforts in conservation. 

Explore the latest edition of AI Magazine and be part of the conversation at our global conference series, Tech & AI LIVE. 

Discover all our upcoming events and secure your tickets today. 

AI Magazine is a BizClik brand